Who is the biggest victim of cheating in the mobile advertising industry? Advertiser! The real money invested only brought short-lived, false prosperity. Today, I will reveal to you the origin, characteristics and prevention measures of cheating in the mobile advertising industry.
A hot mobile advertising market Mobile advertising is considered to be the most active engine of the mobile Internet economy. In 2014, the size of the mobile Internet advertising market soared to 12.5 billion yuan, of which 5 billion yuan came from mobile advertising platforms. In the application market, various types of mobile apps emerge in an endless stream. In order to stand out among the massive applications, promotion has become the most urgent need, and mobile advertising platforms have become the best channel for APP mobile marketing. 2. The generation of cheating behavior Advertisers use "mobile advertising" to push their apps simply to achieve results in rankings (clicks, downloads) and effects (activation, retention). So how does cheating occur? 1. Nutrition from cheating
Since 2013, a large number of online money-making applications have emerged, and there is a strong demand for APP store rankings.
Seeking provides "nutrition" for cheating behavior.
2. The breeding ground for cheating
At the same time, the monitoring technology threshold for mobile advertising is high, the background data structure is complicated, and there are many monitoring dimension indicators, which provide a "breeding ground" for cheating behaviors derived from the mobile advertising process.
Taking the iOS system as an example, in the past, a budget of tens to hundreds of thousands could achieve a good ranking on the list. Later, due to the emergence of fake traffic and cheating, as well as Apple's reduction in weight, the download volume was counted through APP ID, resulting in advertisers not getting the real download volume, so fake traffic ranking did not play a substantial role. Ultimately, the cost of rushing to the list increased greatly. Features of Three Click Cheat According to analysis by industry insiders, click fraud is a relatively low-cost and easy method that can usually be used to send messages directly using a large number of test machines or simulators. Some methods also involve hiring or incentivizing users to make large numbers of clicks. By analyzing the click data logs, it is not difficult to find that click fraud has the following characteristics or phenomena:
1. IP dispersion is dense;
2. Time cycles are repeated;
3. Data is not sent from the mobile terminal but from the simulator;
4. Unable to obtain UA information of the mobile terminal.
The evaluation criteria for effectiveness are mainly reflected in the next-day retention rate, user activity and user recharge behavior. The next-day retention rate used to be relatively high, and most could reach 20%-30%. Later, due to the emergence of cheating behaviors such as brushing, fake traffic emerged in an endless stream, and the conversion rate dropped to 1%.
Effect cheating often occurs in subsequent activation, registration, payment and other behaviors. Common methods include simulating downloads with test machines or simulators that are consistent with clicks, modifying device information through mobile manual or technical means, cracking SDK to send virtual information, and simulating download activation. It is not difficult to find out from the data logs that IP dispersion is dense, time cycles are repeated, simulators are sent, and attribution time differences are illogical.
Advertisers are the biggest victims of industry cheating. It is understood that under the background of increasingly urgent demand for quantifiable monitoring of mobile advertising effectiveness, mobile advertising platforms have already developed and improved a series of proactive and effective anti-cheating mechanisms from the perspectives of advertisers' rights, their own business stability and the sustainable development of the industry. Five measures to prevent cheating Recently, I had an in-depth exchange with an operations expert and learned in detail about some of the current professional mobile advertising platforms’ countermeasures against cheating. The following is some of the useful information that I have sorted out for your reference. 1. Strengthen the SDK anti-cheating module: Provide protection strategies based on abnormal time difference, set the range of time difference of click and activation according to verification analysis, and exclude abnormal data. By updating the new version of SDK to integrate the fraud protection module and adding more device verification information, the SDK will start data channel encryption when transmitting data, increase its cheating cost, and ensure data transmission security.
2. Add device verification information: When receiving an activation request, the server will use a special algorithm to verify each set of data according to the agreed device information, intercept invalid or forged data, and ensure the authenticity and validity of the data.
3. IP dispersion verification function: It provides protection strategies based on abnormal IP segments. Data that exceeds the peak range of the set IP number will be automatically filtered by the anti-cheating function. By retrieving a large number of cheating logs, it is found that in a certain period of time, the excessive concentration of clicks or activated IPs is a common manifestation of data anomalies. The click-through advertising technology system will set the peak value of the number of IPs within a certain period of time. If it exceeds the set range, it will be classified as abnormal data.
4. Advertising effect monitoring + anti-cheating combined function: Any opportunistic behavior will trigger the anti-cheating function system setting, and abnormal data will be filtered out immediately. The mobile advertising fraud process usually occurs in three steps: click, download, and activation. Normally, each step has its corresponding time. During fraud, fake clicks and activations coexist. It doesn’t matter which comes first or later, as long as they happen, so the time logic is often wrong, such as the activation time is earlier than the click time or the time difference is too short. Click-through ads develop a protection strategy based on abnormal time difference, set an abnormal time difference range, and after analyzing the time difference between clicks and activations, verify the time difference setting range and classify devices that activate and convert too quickly as abnormal.
5. With the help of third-party testing agencies: Some third-party monitoring structures on the market rely on their third-party neutrality and take into account the data of the entire platform to provide fraud protection. Finally, I would like to make my usual appeal. Although competition in the mobile advertising market is becoming increasingly fierce, I hope that both advertisers and channels can cultivate their own internal strengths, promote their products in a reliable way, and use our actions to jointly purify the mobile advertising environment. At present, professional mobile advertising platforms/alliances such as Tencent Guangdiantong, Dianru, and Wanpu have very complete anti-cheating mechanisms, which can intercept cheating behaviors through various dimensions (including IP analysis, behavioral habit analysis, system identification, device identification, activation time interval, and click time). If necessary, you can ask for their contact information in the background to discuss how to prevent cheating.
